Glossary
The working vocabulary.
Definitions kept deliberately short. Each term is designed to survive the meeting in which it is used.
- Challenge Path
- The known, safe, reachable route by which anyone affected by an AI-influenced decision can question it.
- Human Point of Consequence
- The named individual who carries the meaning of an outcome and holds the authority to stop the sequence.
- Judgment Abdication
- The pattern in which a human authority delegates a decision to a system without accepting that they still carry its consequence.
- Judgment Apprenticeship
- The deliberate practice of growing the next generation of judgment when routine work is automated.
- Judgment Trail
- The record of what was considered — not merely what was decided — for a consequential AI-influenced decision.
- Space of Autonomy
- The bounded region — inputs, magnitudes, populations, conditions — inside which an AI system may act without human intervention.
- The Second Map
- The working diagram of who actually makes decisions once an AI system enters a workflow.
- Translator
- The person or team that reframes a model's output before a human acts on it. Often invisible, often the most consequential role on the Second Map.
- Trusting Party
- The leader who takes a translated model output as input to their own decision.
- Voice Rights
- The formal authority of a role to be heard, on the record, before a decision closes.